© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 72791
    Prev
    Next
    INDEX
    Explanations

    dates and numbers

    np_acts-logits-general · gemini-2.5-flash-lite

    The neuron strongly activates on four‐digit numeric tokens that represent years (e.g. “2001,” “2014,” etc.).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     permitem
    -1.92
    ≅
    -1.72
     soie
    -1.71
    ˃
    -1.70
     baloncesto
    -1.68
     gustaMe
    -1.62
     laine
    -1.61
    ⩽
    -1.60
     possuem
    -1.60
     ajudam
    -1.59
    POSITIVE LOGITS
    1
    3.03
    4
    2.31
    5
    2.28
    6
    2.25
    7
    2.20
     When
    2.11
    8
    2.05
    3
    2.00
    9
    1.98
     cuando
    1.89
    Activations Density 0.101%

    No Known Activations